Anomaly detection on Hyper-Kvasir
0.972AUCCCD + IGD
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| CCD + IGDImageNet pre-training=true2021.03 | 0.972 | — | — | — | — | |
| CCD + F-anoGANImageNet pre-training=true2021.03 | 0.958 | — | — | — | — | |
| CCD + MS-SSIMImageNet pre-training=true2021.03 | 0.945 | — | — | — | — | |
| IGDImageNet pre-training=true2021.03 | 0.939 | — | — | — | — | |
| DPDLNumber of training anomaly examples=102025.02 | 0.939 | — | — | — | — | |
| CAVGA-RuImageNet pre-training=true2021.03 | 0.928 | — | — | — | — | |
| MS-SSIMImageNet pre-training=true2021.03 | 0.917 | — | — | — | — | |
| ADGANImageNet pre-training=true2021.03 | 0.913 | — | — | — | — | |
| F-anoGANImageNet pre-training=true2021.03 | 0.907 | — | — | — | — | |
| RotNet + IGDImageNet pre-training=true2021.03 | 0.905 | — | — | — | — | |
| AHLNumber of training anomaly examples=102025.02 | 0.88 | — | — | — | — | |
| DRAtraining_anomaly_examples=10, known_anomaly_class=Mean2022.03 | 0.844 | — | — | — | — | |
| DRAC (number of anomaly classes)=4, Training setting=ten random anomaly examples2022.03 | 0.834 | — | — | — | — | |
| DRAtraining_anomaly_examples=10, anomaly_classes=42022.03 | 0.834 | — | — | — | — | |
| DRANumber of training anomaly examples=102025.02 | 0.834 | — | — | — | — | |
| DevNetNumber of training anomaly examples=102025.02 | 0.829 | — | — | — | — | |
| DevNetOpen-set setting=true2021.08 | 0.822 | 0.834 | 0.799 | 0.844 | 0.81 | |
| DPDLNumber of training anomaly examples=12025.02 | 0.821 | — | — | — | — | |
| OCGANImageNet pre-training=true2021.03 | 0.813 | — | — | — | — | |
| FLOSOpen-set setting=true2021.08 | 0.786 | 0.764 | 0.81 | 0.815 | 0.754 | |
| FLOSNumber of training anomaly examples=102025.02 | 0.773 | — | — | — | — | |
| AHLNumber of training anomaly examples=12025.02 | 0.768 | — | — | — | — | |
| DRAtraining_anomaly_examples=1, known_anomaly_class=Mean2022.03 | 0.732 | — | — | — | — | |
| DeepSADC (number of anomaly classes)=4, Training setting=ten random anomaly examples2022.03 | 0.719 | — | — | — | — | |
| DAEImageNet pre-training=true2021.03 | 0.705 | — | — | — | — | |
| DRAtraining_anomaly_examples=1, anomaly_classes=42022.03 | 0.69 | — | — | — | — | |
| DRANumber of training anomaly examples=12025.02 | 0.69 | — | — | — | — | |
| FLOSNumber of training anomaly examples=12025.02 | 0.668 | — | — | — | — | |
| SAOENumber of training anomaly examples=102025.02 | 0.666 | — | — | — | — | |
| DevNetNumber of training anomaly examples=12025.02 | 0.653 | — | — | — | — | |
| MINNSC (number of anomaly classes)=4, Training setting=ten random anomaly examples2022.03 | 0.647 | — | — | — | — | |
| DeepSADOpen-set setting=true2021.08 | 0.63 | 0.666 | 0.672 | 0.619 | 0.564 | |
| MINNSOpen-set setting=true2021.08 | 0.608 | 0.608 | 0.679 | 0.665 | 0.48 | |
| MLEPNumber of training anomaly examples=102025.02 | 0.6 | — | — | — | — | |
| SAOENumber of training anomaly examples=12025.02 | 0.498 | — | — | — | — | |
| MLEPNumber of training anomaly examples=12025.02 | 0.445 | — | — | — | — | |
| KDADOpen-set setting=true2021.08 | 0.403 | 0.405 | 0.404 | 0.435 | 0.367 | |
| KDADtraining_protocol=Unsupervised, known_anomaly_class=Mean2022.03 | 0.403 | — | — | — | — | |
| KDADtraining_anomaly_examples=Unsupervised, anomaly_classes=42022.03 | 0.401 | — | — | — | — |